The dynamic responses of O2 uptake (Vo2) to a range of constant power output levels were related to exercise intensity [as percent maximal Vo2 and as below vs. above lactic acid threshold (LAT)] and to the associated end-exercise lactate in three groups of subjects: group I, untrained subjects performing leg cycle ergometer exercise; group II, the same subjects performing arm cycle exercise; and group III, trained cyclists performing leg cycle ergometer exercise. Responses were described by a double-exponential equation, with each component having an independent time delay, which reduced to a monoexponential description for moderate (below-LAT) exercise. When a second exponential component to the Vo2 response was present, it did not become evident until approximately 80-100 s into exercise. An overall time constant (tau(T), determined as O2 deficit for the total response divided by net end-exercise Vo2) and a primary time constant (tau(P), determined from the O2 deficit and the amplitude for the early primary Vo2 response) were compared. The tau(T) rose with power output and end-exercise lactate levels, but tau(P) was virtually invariant, even at high end-exercise lactate levels. Moreover the gain of the primary exponential component (as DELTAVo2/DELTAW) was constant across power outputs and blood lactate levels, suggesting that the primary Vo2 response reflects a linear system, even at higher power outputs. These results suggest that elevated end-exercise lactate is not associated with any discernible slowing of the primary rise in Vo2. Rather, the increasing overall time constant for Vo2 with increasing power outputs above the LAT is due to the emergence after 80-100 s of a second slowly developing Vo2 component that delays the attainment of a steady state and ultimately increases the O2 cost of the exercise.